Lakehouse | Krishiv

LiveTable

Obtained via session.live_table(name). Provides row-level ingestion into a live SQL-queryable table.

Method	Returns	Description
`name() -> str`	`str`	Return the table name.
`ingest_row(row: dict)`	`None`	Append a single row (dict of column → value).
`refresh()`	`None`	Flush pending inserts into the queryable snapshot.
`change_feed() -> ChangeFeedIter`	`ChangeFeedIter`	Get an async iterator of change records.
`drop()`	`None`	Drop and unregister the live table.

An in-memory Iceberg-like table that supports snapshot-based DML. Useful for testing lakehouse patterns without a real Iceberg catalog.

MemoryLakehouseTable(schema: Schema, name: str = "")

Method	Returns	Description
`append(batches)`	`None`	Append Arrow batches as a new snapshot.
`overwrite(batches)`	`None`	Replace all data with new batches.
`delete_where(predicate: str)`	`int`	Delete rows matching a SQL predicate. Returns deleted count.
`update_where(predicate, assignments)`	`int`	Update matching rows. Returns updated count.
`merge(source_batches, condition, actions)`	`None`	Apply MERGE logic (insert/update/delete) from a source.
`evolve_schema(new_schema)`	`None`	Evolve the table schema (add nullable columns).
`current_snapshot_id() -> int`	`int`	Return the current snapshot ID.

IcebergRestCatalog(uri: str, warehouse: str = None, token: str = None)

Method	Returns	Description
`list_tables(namespace: str) -> list[str]`	`list[str]`	List all table names in a namespace.
`load_table_metadata(namespace, table) -> dict`	`dict`	Load raw Iceberg table metadata JSON.

Function	Description
`read_iceberg(uri, catalog_uri=None) -> DataFrame`	Read an Iceberg table (requires `iceberg` feature).
`read_delta(path, version=None) -> DataFrame`	Read a Delta Lake table directory (requires `delta` feature).
`read_hudi(path, query_type='snapshot') -> DataFrame`	Read a Hudi table.
`write_hudi_append(df, path) -> HudiWriteResult`	Append a DataFrame to a Hudi table.
`write_hudi_upsert(df, path, key_col) -> HudiWriteResult`	Upsert a DataFrame into a Hudi table by key column.

Method	Returns	Description
`instant() -> str`	`str`	Hudi commit instant timestamp.
`rows_inserted() -> int`	`int`	Number of rows inserted.
`rows_updated() -> int`	`int`	Number of rows updated.
`snapshot_rows() -> int`	`int`	Total rows in the table after the write.